Linezolid (LZD) has arisen as an alternative treatment in diabetic foot osteitis due to staphylococci. LZD resistance selection is difficult and involved various molecular mechanisms. As a better knowledge of those mechanisms could be beneficial for pathogenic strains' screening, we simulated in vitro the spontaneous mutagenesis process that leads to LZD-resistant strains from two Staphylococcus epidermidis strains responsible for monomicrobial diabetic foot osteitis. LZD high resistance was selected for both strains, with the same timeline of mutation appearance. Mutation in L3 protein (G152D) occurred first and quickly, but did not cause phenotypically detectable resistance or fitness cost. It was later followed by different 23S rRNA mutations (G2505A, G2447T), leading this time to detectable resistance (minimum inhibitory concentration [MIC] ≥8 mg/L). This phenomenon underlies the difficulty of resistance selection in coagulase-negative staphylococci (CoNS). This study is the first description of G2505A mutation in CoNS. Various phenotypical impacts were observed depending on strain and mutation: (i) fitness cost of G2505A and G2447T mutations; (ii) loss of erythromycin resistance concomitantly with L3 mutation selection; (iii) correlation between number of mutated rrl copies and LZD resistance level for G2447T. In conclusion, the risk of selection of high-level LZD-resistant S. epidermidis strains is weak, but does exist. It could probably appear in case of long-term treatment and be favored in the case of a pre-existing mutation in L3 ribosomal protein. Thus, broad screening conditions for pathogenic strains should probably be considered.
Keywords: 23S RNA; oxazolidinones; ribosome; staphylococci.